WebDAV?

[retog]

Carl, you've mentioned the use of WebDAV. Could you describe in more detail how such an API could look like and how MKCOL, PROPFIND, PROPPATCH and friends could be helpful for the extractor API?

[cblakeley]

Reto,

Use of Virtuoso WebDAV as the proposed mechanism for data import was effectively dropped after the 1st Sprint Review Meeting, as I've indicated in subsequent meetings and emails. The alternative mechanism for data transformation/extraction proposed for Sprint 2 as part of DSS10, which we've been working on subsequently, is for a new stateless Sponger endpoint which supports POSTing of the content to be transformed and which returns the generated RDF directly in the POST response, without saving the data to triple store. This approach was favoured by WP1 to allow chaining of transformers, and fits with SRFG's proposals.

The initial proposal at the Kick Off meeting was to use Virtuoso's WebDAV folders as a mechanism for triggering the Sponger and transformation of any files deposited in these folders to RDF. (Integration with the Sponger would have been through a particular type of Virtuoso WebDAV folder, a "Linked Data Import" folder.)

However at the 1st Sprint Review Meeting when this was demonstrated, issues were raised by WP1 that this approach was problematic

• sponging via this route isn't RESTful, nor completed in a single HTTP request.
• the time taken to transform/sponge the deposited file is not known and completion of the sponging not easily determined
• the sponged data is saved to triple store, so the transformation is not stateless. Consequently we changed our focus for Sprint 2 to the new Sponger endpoint described above.

Since that time, I've been of view that WebDAV does not feature in the envisaged extractor pipeline.

[retog]

Ok, thanks for clarification.